pySigma

pySigma is a python library that parses and converts Sigma rules into queries. It is a replacement for the legacy Sigma toolchain (sigmac) with a much cleaner design and is almost fully tested. Backends for support of conversion into query languages and processing pipelines for transforming rule for log data models are separated into dedicated projects to keep pySigma itself slim and vendor-agnostic. See the Related Projects section below to get an overview.

Getting Started

To start using pySigma, install it using your python package manager of choice. Examples:

pip install pysigma
pipenv install pysigma
poetry add pysigma

Documentation with some usage examples can be found here.

Autodiscovery and Migration of Backend Plugins

Previously, in order to export the objects (backends, pipelines, and validators) of the plugin, it was required to manually export them using global variables in the __init__.py file of the corresponding module. However, this manual export is no longer necessary as pySigma now employs a better mechanism for autodiscovery to locate the exported objects. Consequently, the use of global variables becomes redundant and can be eliminated. Nevertheless, to ensure compatibility with older versions of pySigma and sigma-cli, support for global variables will be retained for the time being. However, it is important to note that in a future release of pySigma, the global variables will be completely phased out.

Pipeline changes after v0.9.11

In the previous implementation, pipelines were defined as functions that returned a ProcessingPipeline object. However, this approach presented a challenge for autodiscovery because the functions lacked type hints, making it difficult to determine their return type. To address this issue, a global variable named pipelines was introduced in the __init__.py file of the plugin's pipeline module. While this workaround allowed the user to manually export the pipeline functions, it was not an ideal solution.

Following discussions here, a decision was made to introduce a class decorator that could be applied to the pipeline functions. This decorator serves two purposes: it allows the functions to be treated as pipeline objects and provides a convenient way for autodiscovery to locate them. Additionally, this approach enables a gradual migration from pipeline functions to pipeline classes without breaking backward compatibility.

By adopting this decorator-based approach, the need for global variables and manual exporting is eliminated. It simplifies the autodiscovery process and enhances code organization. It is important to note that while the previous approach of using global variables will be supported for the time being, it will eventually be phased out in a future release of pySigma.

Here is the revised example illustrating the use of the class decorator:

from sigma.pipelines.base import Pipeline

@Pipeline
def pipeline_1():
    return ProcessingPipeline(
        ... # pipeline code goes here
    )


class Pipeline_2(Pipeline):
    def apply(self):
        return ProcessingPipeline(
            ... # pipeline code goes here
        )

Both pipelines can still be used in the same manner as before. There is no difference between the two approaches because the Pipeline_2 class can be instantiated and used as a pipeline object, like Pipeline_2()() or Pipeline_2().apply(). In other words, when the class is instantiated, an object of the Pipeline class is returned. Calling the object itself will automatically run the apply method, which returns a ProcessingPipeline object. This behavior aligns with the functionality of the pipeline_1 function, which also returns a ProcessingPipeline object. This consistency results in a cleaner and more streamlined approach for autodiscovery and facilitates the gradual migration of pipeline functions to classes.

pySigma before v0.9.11

The backend plugin autodiscovery functionality has been added, eliminating the need for manual registration of plugins in sigma-cli. However, some backends may not function with the updated sigma-cli version. To address this issue, plugin developers should make the following changes to their backends:

1. In the sigma/backends/my_awesome_backend/__init__.py file, add a backends global variable that references the backend class:

   from .my_awesome_backend import MyAwesomeBackend
   
   backends = {
       "my_awesome_backend": MyAwesomeBackend,
   }

2. In the sigma/pipelines/my_awesome_pipelines/__init__.py file, add a pipelines global variable that lists the available pipelines:

   from .my_awesome_pipelines import pipeline_1, pipeline_2
   
   pipelines = {
       "pipeline_1": pipeline_1,
       "pipeline_2": pipeline_2,
   }

3. (Optional) If your backend has Validators (used with sigma check): In the sigma/pipelines/my_awesome_validators/__init__.py file, add a validators global variable that lists the available pipelines:

   validators = {
       "validator_1": MyFirstValidator,
       "validator_2": MySecondValidator,
   }

4. Finally, submit a pull request to the pySigma-plugin-directory and update the version compatibility of your backend plugin with pySigma.

By following these steps, your backend plugin will be compatible with newer versions of pySigma and sigma-cli, allowing for autodiscovery and migration of backend plugins.

Create Your Own Backend for pySigma

The creation of a backend has become much easier with pySigma. We recommend using the "Cookie Cutter Template" and reviewing the existing backends listed in the "Related Projects" section of this README.

pySigma Cookie Cutter Template

Features

pySigma brings a number of additional features compared to sigmac, as well as some changes.

Modifier comparison between pySigma and sigmac

Backends comparison between pySigma and sigmac

On 2022/04/10

Overview

Conversion Overview

Pipelines

More details are described in the documentation.

Testing

pySigma uses pytest as testing framework. Simply run pytest to run all tests. Run pytest --cov=sigma to get a coverage report.

Building

To build your own package run poetry build.

Linting

To lint the code run poetry run black. To check for linting errors run poetry run black --check.

This project also uses pre-commit, which is installed by poetry as part of dev dependencies. To install the git hooks run poetry run pre-commit install after cloning the repository and installing the dependencies.

Contributing

Pull requests are welcome. Please feel free to lodge any issues/PRs as discussion points.

This blog post by Micah Babinski explains the process from a developer's perspective.

Maintainers

The project is currently maintained by:

• Thomas Patzke [email protected]
• frack113

Related Projects

pySigma isn't a monolithic library attempting to support everything but the core. Support for target query languages and log data models is provided by additional packages that extend pySigma:

• sigma-cli: a command line interface for conversion of Sigma rules based on pySigma.
• pySigma-backend-splunk
• pySigma-pipeline-sysmon
• pySigma-pipeline-crowdstrike

All packages can also be installed from PyPI if not mentioned otherwise by the Python package manager of your choice.

License

GNU Lesser General Public License v2.1. For details, please see the full license file located here.